15+ years in marketing. Traditional playbook is retired. I run marketing with AI. I also do marketing for AI. Because agents now buy and people ask AI who to trust before they spend a dollar. The game flipped on both ends. This is where I break it down. Let's #connect

the question used to be "will AI change commerce?" then "how fast?" now it's "what did we miss while we were deciding whether to start?" the answer is always the same: time. the one resource you can't recover. every week of debate was a week of compounding someone else earned.

Shopify AI traffic grew 8x. orders grew 13x. the traffic is converting better than human traffic. and most brands still haven't checked if they show up in the answer that's sending it.

AI visibility work has made me more obsessed with simplicity, but LESS patient with marketing that sounds good and says nothing.

the entire AI visibility discipline in 11 words: make it easy for a machine to recommend you accurately.

The two most useful AI visibility tools cost $0. One shows you what you're sending AI. The other shows you what AI is telling your buyers. 1. Right-click → View Source This shows you what an AI crawler can actually read on your page. Is the price in the HTML? Is the schema correct? Does the first sentence explain what you sell? Or does it just say “Welcome”? That's the input check. What are you feeding the machine? 2. Open ChatGPT in incognito Type your brand name like a stranger would. Then read the answer. What does AI think you sell? Who does it think you're for? What does it get wrong? That's the output check. What is the machine telling your buyers? One shows you what's going in. The other shows you what's coming out. Most brands haven't done either. Not because they're difficult. Because nobody on the marketing team thought to look at the source code. And nobody thought to log out before checking ChatGPT. Two checks. $0 cost. Start there. Everything else is refinement.

There are only 3 types of brands in AI right now. And most brands have no idea which one they are. 1. Invisible AI doesn't mention you. Buyers asking questions in your category never see your name. The fix is straightforward: Add the data. Get crawlable. Get seen. 2. Visible but wrong AI mentions you. But it gets something wrong. Wrong price. Wrong features. Wrong positioning. This is more dangerous than being invisible. Because the buyer sees the answer, trusts it, and acts on it. The fix: Correct the source. 3. Visible and accurate AI describes you correctly. Right product. Right audience. Right price. This is the rarest category. And even here, you're not done. Because accuracy today doesn't guarantee accuracy next month. Sources change. Competitors change. Products change. AI changes. So you have to keep checking. Most brands think they're in category 3. Most are actually in category 2. And they don't know. Because they've never checked what AI says about them.

This B2B SaaS company spent serious money building a website AI couldn't read. Beautiful design. Clean layout. Clear messaging. Compelling product thesis. Feature pages. Use cases. Team section. A great website. Built on a JavaScript rendering framework. No Product schema. No Organization schema. No FAQ markup. No structured data. The site rendered almost entirely client-side. So when an AI crawler arrived, the initial HTML was basically a shell: Script references. And nothing else. We tested 44 non-branded queries across three AI engines. Zero citations from the company's own domain. Not because the content was weak. Because the content was unreachable. The crawler arrives. Sees an empty container. Moves on. We saw something similar in a consumer ecommerce audit. A product page exhausted the AI crawler's token budget on navigation menus before it reached the actual product facts. That brand at least had partial schema, some indexable text, and a crawlable catalog. This one had none of it. The website existed for humans. For AI, it was an empty container with a loading spinner that never resolved. We've seen versions of this across a lot of companies we've audited in the last six months. Beautiful modern websites. Impressive when a human evaluates them. Invisible to the systems increasingly deciding what buyers see. The site looks like it cost $100K to build. To an AI crawler, it looks like: <div id="app"></div> The most expensive blank page on the internet.

The most important content on your website might be the content nobody ever reads. “Content” in 2020 meant blog posts. Articles. Videos. Thought leadership. All still valuable. Still drives backlinks. Still gets cited. Still builds authority. But “content” in 2026 has another layer most teams haven't built yet. Structured data. Schema markup. Comparison tables. FAQs with your brand in the answer. Static HTML pricing. Machine-readable product information. The old content talks to humans. The new layer talks to machines. Both matter. But most brands are missing the second layer entirely. Blog posts build authority. Schema tells AI what your product actually is. The article gets you crawled. The structured data helps determine what AI can extract when it arrives. The strange part? Some of the content humans never read is becoming the content that determines whether machines send buyers to you. And most content teams don't even know this layer exists.

AI is the most honest mirror your brand will ever face. it doesn't care about your intentions. doesn't read your brand guidelines. doesn't attend your all-hands. it shows you exactly how the internet describes you when you're not in the room. most brands don't like what they see. that discomfort is the starting point. not the problem. the problem was always there. AI just showed it to you without being polite about it.

5 AI visibility checks. 10 seconds each. Do them right now. 1. Check your price. Right-click your product page → View Source → Ctrl+F your price. Is it actually in the HTML? If not, AI may not be able to reliably see it. 2. Check your AI crawlers. Go to yourdomain.com/robots.txt. Search for GPTBot or ClaudeBot. See “Disallow”? You may be blocking the engines you're trying to rank in. 3. Check your first sentence. Read the first sentence of your product page. Does it clearly say what you sell, who it's for, and what it does? Or does it say “Welcome”? 4. Check your meta description. Open View Source and find the meta description. Was it written for 2026? Or is it still describing the product you sold in 2021? AI can still use it as a summary signal. 5. Check your feature count. Count the features on your product page. Now ask ChatGPT what your product does. Count the features it lists. Do the numbers match? 50 seconds. At least one of these will probably surprise you. Reply with the number.

Same conversation. Every week. Different client. “We need an AI visibility strategy.” “Have you checked what AI says about you?” “No.” “Let’s do that first.” We check. They go quiet. Then: “How long has it been saying this?” Every time. The check replaces the strategy conversation. Because the strategy is usually the same: Fix what’s wrong. The only variable is what’s wrong. And you can't know that from a meeting. You can only know it by checking. The strategy starts after the screenshot. Never before it.

Give me one day inside your company, and I'll tell you what AI is telling your customers about you. I wouldn't start with a strategy deck. I'd do 6 things. Morning: Type the brand name into 4 AI engines from incognito. Screenshot all 4. Write down every inaccuracy. Then open the main product page. View source. Ctrl+F the price. Check the schema type. Count the actual features vs. what AI thinks the product has. Then check robots.txt. Are the AI crawlers blocked? Lunch: Read the top 3 Reddit and G2 threads about the company. That's part of the corpus AI is reading too. Afternoon: Search: “Best [category] for [audience]” Run it across all 4 engines. Record where the company appears. More importantly, record where it doesn't. 5 PM: Send the 4 brand screenshots to the CEO with one question: “Is this accurate?” That's it. One day. Zero cost. No strategy deck. No six-month roadmap. Just find out what AI already believes about the company. That 5 PM email is where the real work starts.

the highest-converting page on the internet might be a page that doesn't exist yet: your comparison page. a buyer asks "best CRM for remote teams." AI looks for comparison content. finds your competitor's. uses their framing. their pricing. their positioning of you. the buyer never sees your version because your version doesn't exist. you lost the sale to a page you never published. the page that doesn't exist is costing you more than any page that does.

What’s one skill that got MORE expensive because everyone started using AI?

Imagine losing a customer to a company that shut down six months ago. That's happening in AI recommendations. Domain expired. Team dissolved. Product gone. And AI still recommends it. Confidently. Features. Pricing. Use cases. Everything. Because nobody told AI the company died. There's no “brand death certificate.” No universal shutdown notification. AI can keep relying on the last information it found, even after the business is gone. A buyer clicks the recommendation. Gets a parked domain. Or a dead product page. And wonders why the “recommended” company doesn't exist. Meanwhile, your product is live. Updated. Supported. Real. And you're competing for the same AI answer slot against a ghost. That's the strange part of AI visibility: Your competitors aren't necessarily the companies operating today. They're the companies AI still remembers. If you're monitoring AI recommendations, don't just track who's gaining visibility. Track who's being recommended that shouldn't be there anymore.

this space humbles you weekly. the moment you think you've got it figured out, an engine changes its behavior. a crawl picks up something unexpected. a client's perfect site still gets described wrong. the learning never stops because the machine never stops learning. you're optimizing for something that optimizes itself. that's the game. permanent student of a system that never holds still.

ChatGPT describes your product in 3 sentences. Claude uses different words. Perplexity highlights different features. AI Mode mentions a competitor Claude doesn't. same product. same day. same question. 4 engines. 4 descriptions. your buyer only opens one. whichever one they open becomes the truth about your brand. you don't control which engine they pick. but you control the data all 4 engines read. clean the data once. all 4 answers improve. that's the leverage most brands don't realize they have. one source fix. four engine results.

A $2,400/year enterprise product was showing up in AI recommendations as free. The pricing page looked great. Beautiful JavaScript slider. Great UX. But AI couldn't read the actual price. The only price it could find was: “$0 free tier” From a blog post published in 2024. So buyers asking: “What are the best tools under $100?” Saw this enterprise product in the free category. They clicked through. Saw the $2,400 price. Left. Meanwhile, buyers asking: “What are the best enterprise tools?” Didn't see the product at all. Because AI had effectively classified it as free. The fix wasn't rebuilding the pricing page. We added one line of static HTML above the slider: “Plans start at $X/seat/month.” The beautiful slider stayed. The UX stayed. AI finally had one readable number to work with. Your pricing UI can be beautiful for humans and completely invisible to machines. Sometimes the AI visibility fix is just one sentence.

client ran a holiday sale. 40% off. ended January 2. it's September. AI still shows the sale price. the landing page returns a 200 instead of a redirect. cached result still says $47. real price $79. buyers arrive expecting a deal that died 9 months ago. one redirect fixed it.

Some brands will disappear from AI recommendations and won't realize it until the pipeline is already thinner. No dashboard will flash red. No alert will fire. No metric will tell the team: “AI stopped mentioning you three months ago.” Buyers will just quietly stop arriving. Pipeline thins. Deals slow down. Conversion drops. And the team starts looking everywhere else: The market. The economy. Seasonality. The sales cycle. Competitors. Anything except the machine that stopped recommending them. That's the uncomfortable part of AI visibility. You can lose distribution without knowing you've lost distribution. Search rankings give you positions. Ad platforms give you spend and clicks. AI recommendations can disappear quietly. If you're not measuring what AI says about your brand consistently, you might not find out until the revenue impact shows up somewhere else.